Journal of 10.1177/1043986205285055 Morash et al.

Contemporary / Influences Criminal on PoliceJustice Stress

Multilevel Influences on Police Stress

Merry Morash
Michigan State University

Journal of Contemporary Criminal Justice Volume 22 Number 1 February 2006 26-43 2006 Sage Publications 10.1177/1043986205285055 http://ccj.sagepub.com hosted at http://online.sagepub.com

Robin Haarr
UNICEF

Dae-Hoon Kwak
Michigan State University
The prior literature has highlighted a variety of workplace problems, such as racial and gender bias and lack of influence over work activities, as influences on police stress. Additional explanations for police stress include community conditions, for example, high crime rates and size of the community, token status within the police organization, and lack of family and coworker support for work-related activities. In a large-sample, exploratory study, this research examined the workplace problems that were hypothesized to predict stress. It also determined whether community conditions, token status, and lack of social support explained additional variance in officers stress levels. Lack of influence over work activities and bias against ones racial, gender, or ethnic group stood out as important predictors of stress after controls were introduced for demographic variables. Interventions to redesign jobs to afford greater influence and to reduce withindepartment bias are approaches that could reduce police officers stress. Keywords: police; stress; bias; token status; community conditions; gender; race; ethnicity

olice occupational stress is a widespread problem because of its numerous negative effects on individuals and on police organizations. Officers who experience high levels of occupational stress report a high incidence of physical ailments and psychological problems that affect their work performance. Specifically, they commonly have poor health (for reviews, see Cooper & Davidson, 1987; Fletcher, 1988; Kirkcaldy, Cooper, & Ruffalo, 1995), are frequently absent from work (Wright & Saylor, 1991), experience burnout (Brown, Cooper, & Kirkcaldy, 1996; Burke & Deszca, 1986; Crank, Regoli, & Culbertson, 1995; Lord, 1996; Stotland & Pendleton, 1989), are dissatisfied with their jobs (Norvell, Belles, & Hills, 1998), and because of weak organizational commitment, they may not fully invest themselves in their work or they may retire prematurely (Kop, Euwema, & Schaufeli, 1999; Malloy & Mays, 1984; Stevenson, 1988). When individuals are overwhelmed by occupational stress,

26

Morash et al. / Influences on Police Stress 27

they suffer from increased chronic stress, depression, heart disease, stomach disorders, alcohol and drug use and abuse, divorce, and even suicide attempts (Anshel, 2000; Biggam, Power, & MacDonald, 1997; Dietrich, 1989; Lord, 1996; Walker, 1997). It is therefore critical to understand the sources of police occupational stress (i.e., the stressors) and to implement strategies for reducing stressors or, if they cannot be reduced, for assisting officers in coping effectively with them. The policing literature has established that officers are exposed to many sources of stress, and thus law enforcement is a particularly stressful occupation (Alkus & Padesky, 1983; Anshel, 2000; Anshel, Robertson, & Caputi, 1997; Paton & Violanti, 1997; Violanti & Aron, 1994, 1997). Since the mid-1970s, there have been numerous studies of the intensity, prevalence, sources, and effects of police officers stress (Aaron, 2000; Aldag & Brief, 1979; Bannerman, 1997; Brooks & Piquero, 1998; Geick, 1998; Hillgren, Bond, & Jones, 1976; Kroes, Margolis, & Hurrell, 1974; Laufesweiler-Dwyer & Dwyer, 2000; Lord, 1996; Morash & Haarr, 1995; Sewell, 1981; Silbert, 1982; Singleton & Teahan, 1978; Storch & Panzarella, 1996; Stotland & Pendleton, 1989; Teahan, 1975; Terry, 1985; Violanti & Aron, 1993, 1994; Washington, 1981; Wexler & Logan, 1983; White & Marino, 1983). The many studies have clearly documented that an important source of police stress is workplace problems, such as harassment and lack of opportunity to advance. Workplace problems are distinct from other stressors, for example, difficulties in balancing a job and family responsibilities or a persons personality traits and related methods of coping with workplace problems. Workplace problems are troublesome features of the work organization. They include negative interactions with other police officers in the department, feelings about status and opportunities in the department, bias and harassment, and overestimates and underestimates of physical abilities (Morash & Haarr, 1995). For police officers, there is evidence that workplace problems have considerable influence on stress. Wexler and Logan (1983) found in their qualitative research that when women police identified occupational stressors, they highlighted problems with interactions with coworkers. White and Marino (1983) found from their longitudinal, quantitative research on men and women that problems in the organization and with the workplace climate were the only types of stressors that seemed to be causally related to stress. Although some subsequent pieces of research have not found an exclusive influence of workplace problems on stress, workplace problems have repeatedly been shown to be important influences (Brown et al., 1996; Crank et al., 1995; Hurrell, 1995; Juniper, 1996; Kirkcaldy et al., 1995; Morash & Haarr, 1995). A key empirical question is whether token status in an organization and community conditions (e.g., crime rates, size of police department) and lack of support from coworkers and family add to the workplace problem explanation of stress. Specifically, a persons family and extensive support network for work activities might affect stress beyond what can be explained by workplace problems. Also, token racial, ethnic, and gender status in the demographic mix of the department could provide explanation for stress beyond what can be predicted from workplace problems alone.

28 Journal of Contemporary Criminal Justice

Finally, the level of crime and disorder in the area served by the department might explain stress beyond what can be accounted for by workplace problems. To the degree that workplace problems are highly predictive of police officer stress, alleviating those problems would seem to be an effective approach to reducing stress. If support networks, race/ethnicity/gender token status, and the conditions in the broader community served by the police department make independent contributions to stress, specific strategies for prevention and help related to these sources would make sense. If support networks have some independent effect on stress, then efforts could be made to strengthen them. If token status has an influence, then diversification of the police department would be most effective. Findings of the influence of the general characteristics (e.g., amount of crime, extent of poverty) of a police jurisdiction would suggest areas that stress should be addressed in supervision and training to equip officers to effectively manage problems emanating from relatively fixed conditions in the community that the police department serves. This article, therefore, examines the contributions of support network, token status in the organization, and community conditions to explaining stress.

Prior Research on Influences on Police Stress

Workplace Problems
A shortcoming of early workplace and police stress research was that the stressors studied were limited to those identified in the seminal studies (Hillgren et al., 1976; Speilberger, Westberry, Grier, & Greenfield, 1981), and thus they reflect the experiences of male, mostly White, police officers in the 1970s and the 1980s. The resulting research failed to consider distinctive predictors of police stress linked to gender or minority status (Morash & Haarr, 1995). In the mid-1980s, researchers (Ellison & Genz, 1983; Goolkasian, Geddes, & DeJong, 1985; Wexler & Logan, 1983) began to give some attention to female officers stress and the unique female-related stressors, including disapproval from fellow officers, coworkersdemands that women prove themselves, and a lack of informal tutoring or mentoring. In the 1990s, a study of 1,087 police officers from 24 police departments across the United States (Morash & Haarr, 1995) considered an array of workplace problems that original fieldwork and other prior research had identified for diverse groups of police officers. Building on a qualitative study of women police and the prior literature, the study identified the workplace problems: overestimate and underestimate of physical abilities, perceived lack of advancement opportunity, ridicule and being set up by other officers, perceived lack of influence on how police work is accomplished, feeling invisible in the department, language harassment, sexual harassment, and stigmatization because of appearance.1 Racial, gender, and Hispanic versus non-Hispanic groups did experience some of the same workplace problems, but as might be expected, some problems were more common among racial minority groups, Hispanics, and women. The range of workplace problems identified and studied by Morash and Haarr (1995) was considered in this research.

Morash et al. / Influences on Police Stress 29

Support Networks
There is evidence that some stress emanates not from an individuals workplace problems but from the lack of support from networks at work and at home. Support networks include family, superiors at work, and peers at work. Each of these groups might provide some understanding of and assistance to police officers with regard to issues related to work. In policing and in other occupations, lack of social support can fail to ameliorate or can exacerbate the effects of workplace problems on psychological and physical health (Cullen, Lemming, Link, & Wozniak, 1985; LaRocco, House, & French, 1980).

Token Status in the Demographic Mix of the Department

An individuals racial, ethnic, and gender minority status within a department affects her or his standing as a token, and this might contribute to stress. Pertinent to token status, in her seminal work, Kanter (1977) found that in various types of organizations, token racial or gender status resulted in stress (pp. 239-240). More recently, Krimmel and Gormley (2003) discovered that women in departments with less than 15% female officers reported low levels of job satisfaction (also see Belknap & Shelly, 1992), high levels of job-related depression, and low self-esteem. Token status can influence acceptance by majority officers (Ellison & Genz, 1983, p. 68), lack of role models (Wexler & Logan, 1983), and outright bias (Ellison & Genz, 1983; Goolkasian et al., 1985).

Community Conditions and Organizational Features

The broader context of policing includes indicators of life quality and social disorder, for example, percentage poor in the area served by the police department, the amount of crime in the area, and metropolitan versus county setting. Highly correlated with these factors are the structure and the size of the police department. In general, large municipal departments (as opposed to county sheriff departments)2 are found in the urban areas where at least some areas are marked by high density, high levels of poverty, and high levels of crime and disorder.3 Officers in small departments that serve small cities, suburban, and rural areas do have different behaviors and attitudes from those who work in large urban and metropolitan centers (Brown, 1981; Mastrofski, Ritti, & Hoffmaster, 1987; Meyers, Heeron, Hingson, & Kovonock, 1987; Powell, 1990). The more militaristic, bureaucratic, discipline-oriented nature of large urban departments may result in stressors beyond individually perceived workplace problems. Consistent with this possibility, Gaines and his coauthors (Gaines, Southerland, & Angell, 1991) reasoned that officers in large departments were stressed because they thought the agency was selfserving and unresponsive. Also, in large departments, managers and executives might lose control, and this could contribute to stress (Regoli, Crank, & Culbertson, 1989). Finally, more informal relationships between community members and officers and

30 Journal of Contemporary Criminal Justice

between superiors and subordinates might lead to less stress in smaller agencies (Regoli et al., 1989).

Statement of the Research Question

Much of the research literature has considered various workplace problems as experienced by individuals as key predictors of stress in policing. This study first considered whether a variety of workplace problems did indeed explain stress. Furthermore, it addressed the question of whether support networks, community characteristics, and a persons status as a token in a department would explain additional variance in stress. Because an officers gender, minority status, education, age, and social class background might alter the predictors of stress, these were used as control variables.

Method
Sample
The study used data from research that was conducted to compare officer responses in departments that had participated in prior research in 1990. In 1990, 24 departments were recruited for research through advertisement among the members of the Police Executive Research Forum, an organization that includes police departments that are interested in applying the results of research. Of the 24 departments that were initially surveyed, 11 agreed to participate in 2003. Police departments were requested to ask 30 individuals in each of the following groups to complete the survey: African American women, African American men, Asian women, Asian men, female Hispanics, male Hispanics, White women, and White men. In many departments, there were not 30 individuals in each category, so a smaller group was approached. In 2003, 947 police officers from the 11 police departments returned the survey, and there was a substantial mix of gender and racial groups and of both Hispanics and non-Hispanics. We worked individually with each department to develop a sampling plan that would approximate a representative sample for groups of interest. In some cases, particularly in small departments, it was possible to actually randomly sample officers from a stratified list. In other departments, after we ascertained that officers who reported for different shifts were generally representative of officers in the department, individuals who reported for a certain shift were approached. One of the researchers spoke with a contact person in each department and negotiated an approach to sampling that would result in a group that would be reasonably representative but that the department felt was feasible given its workload and resources. We encouraged departments to allow work time to complete the survey, but a few departments were only willing to distribute the survey by internal mail and ask that they be brought back. These departments had the lowest response rates. In all, 2,051 individuals were asked to take part in the survey, and 947 (46.2%) returned a survey.

Morash et al. / Influences on Police Stress 31

The 11 departments that participated in the research included an East Coast department that served a population of more than one and a half million, a Southwestern department that served a population of about one-half million, and a Midwestern department serving a population of slightly less than 400,000. The departments that served medium-sized areas, ranging from 130,000 to 34,000 population, were in rural/ small city/suburban areas of Maryland, North Carolina (2), Iowa, and California. Departments serving populations of less than 100,000 were in Georgia, Oregon, and Florida. Although we oversampled racial minority and female officers, European American males were still a large proportion of the sample, because many departments did not have the desired number of employees in minority gender, racial, or ethnic groups. Males made up slightly more than 72% of the sample, whereas females represent 27% of the sample. In addition, two thirds of the officers were Caucasian, and racial minority officers made up one third. The sample did not represent the actual proportions of women and men, racial minority and racial majority officers in the departments, but instead was designed to allow for inclusion of a broad range of gender and race groups and Hispanics and non-Hispanics. The final sample of departments and individuals in the departments was not intended to be a probability sample. However, the large sample size gives us some confidence that the various subgroups were reasonably well represented. The sample allows exploration of predictors of stress. The research, therefore, is able to shed some light on hypotheses that might be most fruitfully tested in further research that does examine probability samples.

Variables and Measurement

Workplace problems. Items developed and validated by Morash and Haarr (1995) were the indicators of workplace problems. Morash and Haarr based their items on the literature on stress in policing, which before their study had focused primarily on White, male police officers, on the limited amount of research on police women and minority race police, and on their own qualitative research on police women. Specific problems that were measured were overestimates of physical ability, underestimates of physical ability, perceived lack of advancement opportunity, ridicule and setups, lack of influence, invisibility, language harassment, bias, sexual harassment, and stigmatization because of physical appearance. Reliability coefficients were recalculated for the scales, and they ranged from .60 to .92 (see Appendix A). Only 2 of the 10 coefficients were below .70. These workplace problems have been documented as important across gender, racial, and ethnic subgroups (Ellison & Genz, 1983, p. 64; White & Marino, 1983). Support networks. Scales to measure support networks were adapted from Cullen et al. (1985). Separate scales had been developed for work-related support from family, from coworkers, and from superiors. Our research showed that support levels from

32 Journal of Contemporary Criminal Justice

peer-level coworkers and superiors were highly correlated, so they were included in one scale, support from coworkers. The reliability was .73. The reliability for the fam4 ily support scale was also acceptable at .74. Token status. Separate indicators were used to reflect token status based on race, gender, and ethnicity (Hispanic, non-Hispanic). For each, the first step was to calculate the percentage of people in the same demographic category as the respondent. For example, for an individual who was White, the number would reflect the proportion of individuals in the department who were White. Prior research has suggested that an individual holds token status when she or he is from a demographic group that includes 15% or less of the people in the organization (Krimmel & Gormley, 2003, pp. 75-76).5 Additional data to determine token status were taken from the Law Enforcement Management and Administrative Statistics (LEMAS) for 2000. (Two departments were not included in the LEMAS database.) Token status was coded so that 1 indicated that the person was in a department where 15% or fewer of the officers shared the particular demographic characteristic, and 0 indicated that the person was not in a 15% token group within the department. Community conditions. The community data were obtained from LEMAS (Bureau of Justice Statistics, 2000), the 2000 Census, and the Uniform Crime Reports for 2000. Due to multicollinearity, it was problematic to untangle urban setting, department size, high crime rates, percentage of the population that is below the poverty line, ratio of police to the population, and other indicators of stressed urban environments. Other researchers have confronted the same difficulty. For example, Brooks and Piquero (1998) concluded from their research that although large department size was predictive of high stress, it might not be the department size, per se, that accounted for the finding. Larger departments are typically serving at least some high crime neighborhoods (Crank & Caldero, 1991), and thus high crime rates might account for the association with stress. Initially, we selected the following variables as indicators of the type of community served by the department: percentage poor, percentage minority, municipal versus county department, violent crime rate, property crime rate, unemployment rate, population density, and ratio of the number of officers to the population.6 We used the following strategy to identify variables that did not create an unacceptably high level of multicollinearity. We began with the assumption that high violent crime rate would be related to stress. From there, we added each variable that did not increase the Variance Inflation Factor (VIF) beyond an acceptable level. In addition to violent crime rate, the variables included in the analysis were property crime rate and type of agency (coded as 1 = municipal, 0 = county). Based on the intercorrelations, we can conclude that for our sample, areas with a high violent crime rate and a municipal police department also had a high percentage of property crime, a high percentage of households below the poverty line, a high percentage of minority residents, high density, and a high ratio of police to population.

Morash et al. / Influences on Police Stress 33

Stress. Responses to four agree-disagree items were combined to indicate selfreported occupational stress for the past year:
The amount of unwanted stress on my job has had a negative effect on my physical well-being. I really felt a lot of unwanted emotional stress from this job. My feeling is that I needed to get some special help in managing the stress of my job. It seems that I can deal with the tensions of my job to the point that they do not interfere with family and social life (reverse coded).

These four items are measured using a Likert-type scale ranging from 1 (strongly agree) to 5 (strongly disagree). The stress scale was computed by adding up the values of each item and then dividing by number of items (M = 2.66, SD = .86). The Cronbachs alpha coefficient ( = .78) indicated an acceptable reliability. The distribution of the measure of stress was somewhat positively skewed. Thus, although the mean accurately reflects that many officers did not report much stress, it should be noted that more than one quarter of officers reported some stress, and a substantial number (12%) indicated high agreement with all of the statements indicating stress. Control variables. The analysis did include five control variables that might account for interconnections of stress with the workplace problems, social support, token status, and nature of the community served by the police department. Gender was coded as 1 for female and 0 for male. Education was coded as 1 for bachelors degree or higher and 0 for less than a bachelors degree. Age was collected on an ordi7 nal scale ranging from younger than 20 years to older than 60 years. For the measure of minority status, White non-Hispanics were coded as 0, and all other individuals were coded as 1. The measure of social class when growing up was a self-report ordinal scale.

Analysis
Ordinary least squares (OLS) regression analysis was used to determine the additional variance explained by organization and community condition variables beyond what could be explained by workplace problems. Thus, at Step 1, all of the workplace problem indicators were entered. At Step 2, all of the indicators of organizational and community conditions were entered. In the final step, demographic characteristics of the individual were added as control variables. Analysis included a test of the significance of the overall R2 at each step, a test for the significance of the increase in R2 at Steps 2 and 3, and tests of the significance for coefficients for independent variables.

Findings
Descriptive statistics are included in Table 1. Regression diagnostics were carried out to examine potential problems with multicollinearity. Based on the findings, mea-

34 Journal of Contemporary Criminal Justice

Table 1 Descriptive Statistics

M Stress scale Stressorsindividual level Physical overestimate Physical underestimate Lack advance. opportunity Ridicule and setups Lacks influence Invisible Bias Language harassment Sexual harassment Stigma and appearance Stressorslack of support, token status, and community conditions Family support Work group support Violent crime rate Property crime rate Token sex Token race Token ethnicity Type of agency County/sheriff = 0 Municipal police dept. = 1 Control variables Sex Female = 1 Minority Minority = 1 Education Less than bachelors degree = 0 Bachelors degree or higher = 1 Age Class 2.66 2.41 2.64 2.53 1.98 3.18 2.22 2.46 1.35 1.06 1.16 SD .86 .66 .82 .89 .56 .87 .62 .79 .70 .28 .45 N 931 936 940 924 920 926 929 912 918 910 918 %

3.86 6.71 797.78 5,517.17 .14 .11 .12

.75 1.26 498.17 1,995.00 .35 .31 .32

921 910 947 947 819 817 813 202 653 23.6 76.4

241 169 577 332 910 913

26.5 18.7 63.5 36.5

3.13 2.48

.93 .75

sures of support from peers at work and support from superiors at work were combined into an indicator of work group support. After that change was made, there were no further problems with multicollinearity. Appendix B includes the correlation matrix for independent and control variables. In the regression analysis, the workplace problems explained 34% of the variance in stress, and the finding was statistically significant (see Table 2). The additional variance explained by the social support, token status, and community indicators was just 2.8%. Although the increase in explained variance is small, it is statistically significant at the .05 level (F = 3.4, df = 9,706). The additional variance (.9%) explained by the

Morash et al. / Influences on Police Stress 35

Table 2 Results of the Regression of Stress on Workplace Problems, Social Support, Community Conditions, Token Status, and Control Variables
Model 1 r Workplace problems Physical overestimated Physical underestimated Lack advance. opportunity Ridicule and setups Lacks influence Invisible Bias Language harassment Sexual harassment Stigma and appearance Support network, community conditions, and token status Family support Work group support Violent crime rate Property crime rate Token sex Token race Token ethnicity Type of agency Control variables Sex Minority Education Age Class when growing up R2 adjusted df F *p .01. **p .05. b r Model 2 b r Model 3 b

.24* .33* .33* .41* .35* .40* .49* .26* .12* .24*

.14 .09 .01 .14 .17 .08 .31 .10 .26 .19

.10* .08** .01 .09** .18* .06 .29* .08** .08** .09**

.13 .05 .04 .17 .17 .08 .30 .13 .27 .17

.10* .05 .04 .11* .16* .06 .27* .10* .09** .09**

.12 .05 .03 .18 .18 .08 .31 .13 .25 .15

.09* .05 .03 .12* .18* .06 .29* .10* .08** .08

.23* .10 .32* .02 .05 .00 .12* .00 .11* .22 .07** .22 .03 .07 .06 .06

.08* .04 .02 .11** .09* .07** .03 .03 .14* .05 .03 .03 .05 .38

.10 .03 .00 .00 .27 .04 .05 .03 .04 .29 .08 .05 .01 23, 687 19.9*

.08* .04 .06 .12** .11** .01 .02 .02 .02 .13* .04 .06 .01

.34 10, 836 44.5*

.37 18, 707 24.8*

addition of the control variables was not statistically significant (F = 2.0, df = 5,701, p > .05). The findings about which workplace problems were the most predictive of stress remained very stable with the addition of token status, community variables, social support, and control variables. In the final step of the regression analysis, dealing with bias among coworkers was most predictive of stress when other variables were controlled. Officers reporting high stress said they felt stress from racial or ethnic bias and they spent considerable time and energy dealing with and helping other officers deal with prejudice and bias. The second strongest predictor of stress was feeling a lack of

36 Journal of Contemporary Criminal Justice

influence on how police work is accomplished. Officers who felt stressed said they could not influence the way policing was done and could not influence department policies and procedures.

Discussion and Conclusion

Our findings are consistent with Schaubroecks (1999) argument that organizational psychologists (and we would add others who are concerned about levels of workplace stress) should pay attention to objective workplace conditions that need to be changed in order to reduce employee stress. He also pointed to the inattention to the construct of job control, which has a demonstrated effect on stress (e.g., Fox, Dwyer, & Ganster, 1993). In his view, explanations that do not highlight workplace conditions but that give prominence to attributions of the cause of stressors and resulting emotions (e.g., Perrewe & Zellars, 1999) suggest strategies that will try to change police officers who are stressed rather than behaviors of their peers, their superiors, or other workplace problems that cause stress. Although job control is a different construct from what we have measured, it is quite similar in meaning to our measure of influence over how the job of policing is accomplished. The primary implication from our research is that root causes of stress might be most fruitfully addressed by attacking problems such as bias and lack of job control or influence over ones work. In interpreting findings about community conditions (e.g., crime rates) and organizational features, such as whether the department serves a county or a city, it is crucial to keep in mind that because of multicollinearity in our sample, in this study, the high crime rates and municipal type of agency stand as reflections of urban, high poverty, high density, and high minority concentration areas. It is somewhat heartening that relatively intractable features of the community, such as crime rates and poverty, do not appear to have strong predictive value in explaining stress. This is not to say that reducing bias and increasing influence on the job are easy to achieve. Full implementation of community policing might be a way to allow officers greater discretion and control of their work, and many departments have been able to implement this approach to policing either for some or, in a few cases, all types of officers (Morash & Ford, 2002). However, the move to community policing, other approaches to giving police more influence over their work, and reductions of bias among officers are not easily changed and would require considerable attention before they could be reduced to the point that stress would decrease. In our analysis of token status as the only or one of a small proportion of a gender, race, or ethnic group, only token status as a female has a significant effect on stress, and neither the correlation nor the beta coefficient suggested that this connection was particularly strong. This set of findings does not let us off the hook for further diversifying police departments, however. Diversification of police departments has many benefits besides potential effects on stress. Also, it is entirely possible that measures of bias, stigmatization, and exclusion explained much of the stress that results from token status. Moreover, additional research should compare the influence of multiple token

Morash et al. / Influences on Police Stress 37

statuses on stress. Other research with the same data set has shown that the individuals who have the most severe workplace problems are African American women (Morash, Haarr, & Gonyea, forthcoming). The intersection between token statuses based on race, gender, and ethnicity can have an influence on individuals above and beyond the effects of any one of these markers of a token status within an organization. It was unexpected that a high property crime rate would be related to low levels of stress. It may be that communities with a high property crime rate have more to steal because they have a relatively high level of resources. Thus, the generally higher resources in a community might offset any stress related to property crime rates. Also unexpected, the violent crime rate was unrelated to stress levels. It may be that even in a high crime jurisdiction, many police officers are not directly dealing with violence much of the time. Crime tends to be concentrated in hot spots within most cities. Or, it also is possible that crime rate is not a primary stressor for police. Indeed, they may self-select into police work because they have the capacity to cope with violence and disorder. An important limitation of this research should be noted. The departments are not representative, and the samples within those departments were not random. Research to replicate the research in a wide variety of police organizations would be very useful. However, the samples are relatively large and a diversity of officers is included. Also, the departments were quite diverse from each other, including some in large urban centers, some in smaller cities, and some in counties that included substantial rural areas. To the extent that police organizational culture and the structure and practices of police departments create stressors, efforts aimed at how officers treat each other and at their control over work activities would be most effective. Efforts to reduce the predictors of stress must be coupled with facilitation of officers use of effective coping strategies. However, even if officers use the most effective strategies for coping with workplace problems, police departments bare the burden of reducing or eliminating the workplace conditions that contribute to stress.

38

Appendix A Scales Measuring Stress and Stressors

Item Range

Scale

Dependent variable Stress 1. Stress has a negative effect on physical well-being 2. Feel a lot of unwanted emotional stress 3. Needed to get some special help (stress) 4. Can deal with tensions 1 to 5 1 to 5 1 to 5 1 to 5

.78

Independent variables Physical ability overestimate

.63 .60 .73

Physical ability underestimate

Lack of advancement opportunity

Ridicule and setups

1 to 4.8

.73

Lack of influence

1 to 5

.75

Invisible

1. Can do more than my physical abilities permit 2. Superiors overestimate my physical capabilities 1. Superiors underestimate physical capabilities 2. Do not realize my physical abilities 1. Opportunity for promotion 2. Apply for and get assigned to the jobs 3. Less a chance to advance 1. Dangerous risks 2. Coworkers give encouragement 3. Superiors set me up to make mistakes 4. Coworkers ridiculed me when ask how to do the job 5. Superiors ridiculed me when ask how to do the job 1. Lack of influence 2. Do not have the power to change the way police work 3. Influence on department policies and procedures 4. Feel that I have a say in how things get done 1. Coworkers forget me 2. Coworkers do not look me in the eye 3. Superiors do not look me in the eye 4. Superiors forget me 5. Coworkers did not thank

1 to 5

.77

Language harassment

1 to 5

.81

Bias

1 to 5

.79

Sexual harassment

1 to 4.9

.92

Stigma and appearance

1 to 5

.84

Family support

1 to 5

.74

Work group support

1. Superiors use profane language 2. Coworkers use profane language 3. Superiors joke about sex 1. Do not feel any stress from racial or ethnic bias 2. No bias against people of my sex, age, race, or ethnic group 3. I spent time/energy dealing with prejudice and bias 4. I spent time/energy helping other police officers deal with bias 5. Dealing with other officers bias uses up a lot of my energy 1. Superiors try to have a romantic type of relationship with me 2. Coworkers viewed pornographic films or photos 3. Coworkers physically touch me 4. Coworkers make jokes about how good-looking I am 5. Superiors force me to have sexual intercourse 6. Superiors make jokes about how good-looking I am 7. Superiors make unwanted comments (homosexual) 8. Coworkers make unwanted comments (homosexual) 1. Superiors joke or comment about my physical size 2. Coworkers joke or comment about my physical size 3. Coworkers make unwelcome jokes about how unattractive I am 4. Superiors make unwelcome jokes about how unattractive I am 1. Get the support I need to feel better 2. No one (in family) can talk about job 3. Support spouse or steady mate 4. Had people who talk about any problems 1. Superiors give encouragement (to do best job) 2. Superiors give encouragement (did job well) 3. Superiors stressed the importance of job 4. Superiors give encouragement (think of better way) 5. Fellow officers give encouragement 6. Think of better ways of getting the work done 7. Fellow officers compliment 8. Fellow officers spent hardly any time helping me 9. Back me up when I make mistakes 1 to 5 .73

39

40

Appendix B Correlation Matrix for Independent and Control Variables

(1) 1 .21* 1 .13* .31* 1 .24* .40* .36* .08** .26* .46* .28* .42* .39* .19* .32* .42* .05 .17* .16* .13* .14* .09* .13* .20* .16* .15* .17* .21* .11* .27* .46* .03 .01 .10* .03 .04 .03 .05 .06 .01 .03 .06 .01 .05 .07** .00 .05 .03 .05 .00 .08**.04 .00 .01 .00 .00* .03 .01 .12* .07** .06 .01 .01 .06 1 .31* 1 .62* .30* 1 .46* .30* .46* 1 .27* .15* .23* .28* 1 .28* .04 .23* .18* .44* 1 .33* .11* .27* .21* .48* .58* 1 .20* .12* .28* .17* .08**.06 .12* 1 .41* .46* .47* .34* .26* .14* .21* .27* .03 .05 .07** .19* .07** .07** .00 .04 .02 .05 .00 .03 .05 .04 .08** .02 .00 .02 .04 .08** .07 .03 .00 .03 .02 .04 .02 .03 .04 .03 .07 .01 .02 .01 .03 .06 .03 .03 .05 .01 .02 .03 .02 .08** .04 .01 .02 .00 .04 .07** .03 .20* .16* .13* .04 .04 .02 .05 .08** .15* .05 .06 .00 .04 .01 .04 .04 .03 .02 .02 .01 .02 .07**.08** .02 .06 .04 .01 .02 .09* .05 .07**.05 .06 .01 .03 .02 .05 1 .06 1 .03 .30* .02 .20* .01 .05 .00 .22* .07 .70* .05 .23* .04 .33* .04 .29* .11* .02 .08**.10* 1 .04 1 .09**.03 .07 .10* .73* .10* .04 .67* .07**.10* .11* .18* .10* .00 .00 .05 (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23)

1. Phy. Ovr 2. Phy. Und. 3. Lack. Adv. 4. Rid. Setup 5. Lack. Inf. 6. Invisible 7. Bias 8. Lan. Hara. 9. Sex. Hara. 10. Stig. App. 11. Fam. Sup. 12. Work. Sup 13. Vio. Crime 14. Pro. Crime 15. Token Sex 16. Token Race 17. Token Ethn. 18. Type. Agen. 19. Sex 20. Minority 21. Education 22. Age 23. Class

1 .06 1 .02 .12* 1 .03 .08** .09** 1 .67* .11* .12* .12* 1 .03 .17* .26* .02 .07** 1 .01 .03 .05 .06 .01 .07* 1 .03 .12 .10 .01 .08** .16* .10*

*p .01. **p .05.

Morash et al. / Influences on Police Stress 41

Notes
1. Job control, which is the extent to which people have discretion and choice in their work, is a construct that is similar to influence over how one gets the job done, which we have measured. In the field of occupational health, job control has been extensively studied and confirmed as an influence on stress (Bond & Bunce, 2001; Terry & Jimmieson, 1999). 2. We used the LEMAS designation of departments as either municipal or county. 3. It should be noted that these correlations apply to our sample of departments and not necessarily to all departments in the United States. For example, some county departments serve quite urbanized areas. However, in our sample, county departments served rural and suburban areas with low population density. 4. Cullen et al. (1985) initially measured peer, supervisory, family, and community support. The alpha values for their scales ranged from .74 to .84. 5. Some research has used the cutoff point of 25%. We calculated alternative indicators of sex, minority, and ethnic group token status using this cutoff, but the results were not markedly different from when the 15% cutoff was used. 6. The violent crime rate is the number of reported violent offenses per 100,000 people in the population. Similarly, the property crime rate is the number of reported property offenses per 100,000 people in the population. 7. Years of employment as a police officer was initially considered, but it was highly correlated with age.

References
Aaron, J. (2000). Stress and coping in police officers. Police Quarterly, 3, 428-450. Aldag, R. J., & Brief, A. P. (1979). Examination of a measure of higher-order need strength. Human Relations, 32, 705-718. Alkus, S., & Padesky, C. (1983). Special problems of police officers: Stress related issues and interventions. Counseling Psychologist, 11, 55-64. Anshel, M. H. (2000). A conceptual model and implications for coping with stressful events in police work. Criminal Justice and Behavior, 27, 375-400. Anshel, M. H., Robertson, M., & Caputi, P. (1997). Sources of acute stress and their appraisals and reappraisals among Australian police as a function of previous experience. Journal of Occupation and Organizational Psychology, 70, 525-549. Bannerman, E. D. (1997). Female police officers: The relationship between social support, interactional style, and occupational stress and strain. Vancouver: Simon Frasier University. Belknap, J., & Shelly, J. K. (1992). The new lone ranger: Police women on patrol. American Journal of Police, 12(3), 47-73. Biggam, F. H., Power, K. G., & MacDonald, R. R. (1997). Self-perceived occupational stress and distress in a Scottish police force. Work and Stress, 11, 118-133. Bond, F. W., & Bunce, D. (2001). Job control mediates change in a work reorganization intervention for stress reduction. Journal of Occupational Health Psychology, 6, 290-302. Brooks, L. W., & Piquero, N. L. (1998). Police stress: Does department size matter? Policing: An International Journal of Police Strategies & Management, 21(4), 600-617. Brown, J., Cooper, C. L., & Kirkcaldy, B. (1996). Occupational stress among senior police officers. British Journal of Psychology, 87, 31-45. Brown, M. K. (1981). Working the street: Police discretion and the dilemmas of reform. Troy, NY: Russell Sage Foundation. Bureau of Justice Statistics. (2000). Law Enforcement Management and Administrative Statistics, 2000. Retrieved December 7, 2005, from http://www.ojp.usdoj.gov/bjs/abstract/lemas00.htm Burke, R. J., & Deszca, E. (1986). Correlates of psychological burnout phases among police officers. Human Relations, 39, 487-502.

42 Journal of Contemporary Criminal Justice Cooper, C. L., & Davidson, M. (1987). Sources of stress at work and their relation to stressors in non-working environments. In R. Kalimo, M. A. El-Batawi, & C. L. Cooper (Eds.), Psychological factors at work and their relation to health (pp. 99-123). Geneva: World Health Organization. Crank, J. P., & Caldero, M. (1991). The production of occupational stress in medium-sized police agencies: A survey of line officers in eight municipal departments. Journal of Criminal Justice, 19, 339-359. Crank, J., Regoli, R., & Culbertson, R. (1995). Institutional and organizational antecedents of role stress, work alienation, and anomie among police executives. Criminal Justice and Behavior, 22, 152-167. Cullen, F. T., Lemming, T., Link, B. G., & Wozniak, J. F. (1985). The impact of social supports on police stress. Criminology, 3, 503-522. Dietrich, J. (1989). Helping subordinates face stress. Police Chief, 56, 44-47. Ellison, K. W., & Genz, J. (1983). Stress and the police officer. Springfield, IL: Charles C Thomas. Fletcher, B. C. (1988). The epidemiology of occupational stress. In C. L. Cooper & R. Payne (Eds.), Causes, coping and consequences of stress at work (pp. 12-57). Chichester, UK: Wiley. Fox, M. L., Dwyer, D. J., & Ganster, M. L. (1993). Effects of stressful job demands and control on physiological and attitudinal outcomes in a hospital setting. Academy of Management Journal, 36, 289-318. Gaines, L., Southerland, M., & Angell, J. (1991). Police administration. New York: McGraw-Hill. Geick, E. (1998). Occupational stress of female police officers: An empirical investigation, Texas A&M University. Goolkasian, G. A., Geddes, R. W., & DeJong, W. (1985). Coping with police stress. Washington, DC: Government Printing Office. Hillgren, J. S., Bond, R., & Jones, S. (1976). Primary stressors in police administration and law enforcement. Journal of Police Science and Administration, 4, 445-449. Hurrell, J. J. (1995). Police work, occupational stress, and individual coping. Journal of Organizational Behavior, 16, 27-34. Juniper, D. (1996). Stress in a police force. Police Journal, 69, 61-69. Kanter, R. (1977). Men and women of the corporation. New York: Basic Books. Kirkcaldy, B., Cooper, C. L., & Ruffalo, P. (1995). Work stress and health in a sample of U.S. police. Psychological Reports, 76, 700-702. Kop, N., Euwema, M., & Schaufeli, W. (1999). Burnout, job stress, and violent behaviour among Dutch police officers. Work and Stress, 13, 326-340. Krimmel, J. T., & Gormley, P. E. (2003). Tokenism and job satisfaction for policewomen. American Journal of Criminal Justice, 28(1), 73-88. Kroes, W. H., Margolis, B. L., & Hurrell, J. J., Jr. (1974). Job stress in policemen. Journal of Police Science and Administration, 2, 145-155. LaRocco, J. M., House, J. W., & French, J.R.P., Jr. (1980). Social support, occupational stress, and health. Journal of Health and Social Behavior, 21, 202-218. Laufesweiler-Dwyer, D. L., & Dwyer, G. R. (2000). Profiling those impacted by organizational stressors at the macro, intermediate, and micro levels of several police agencies. Justice Professional, 12, 443-469. Lord, V. B. (1996). An impact of community policing: Reported stressors, social support, and strain among police officers in a changing police department. Journal of Criminal Justice, 24, 503-522. Malloy, T. E., & Mays, L. (1984). The police stress hypothesis: A critical evaluation. Criminal Justice and Behavior, 11, 197-224. Mastrofski, S., Ritti, R., & Hoffmaster, D. (1987). Organizational determinants of police discretion: The case of drinking-driving. Journal of Criminal Justice, 15, 387-402. Meyers, A. R., Heeron, T., Hingson, R., & Kovonock, D. (1987). Cops and drivers: Police discretion and the enforcement of Maines OUI Law. Journal of Criminal Justice, 15, 361-368. Morash, M., & Ford, J. K. (2002). The move to community policing: Making change happen. Thousand Oaks, CA: Sage. Morash, M., & Haarr, R. N. (1995). Gender, workplace problems, and stress in policing. Justice Quarterly, 12(1), 113-140. Morash, M., Haarr, R. N., & Gonyea, D. P. (forthcoming). Workplace problems in police departments and methods of coping: Women at the intersection. In L. Goodstein, C. Renzetti, & S. Miller (Eds.), Women, crime, and criminal justice. Los Angeles: Roxbury Press.

Morash et al. / Influences on Police Stress 43 Norvell, N., Belles, D., & Hills, H. (1998). Perceived stress levels and physical symptoms in supervisory law enforcement personnel. Journal of Police Science and Administration, 16, 75-79. Paton, D., & Violanti, J. M. (1997). Long-term exposure to stress and trauma: Addiction and separation issues in police officers. In G. M. Hubermann (Ed.), Looking back and moving forward: 50 years of New Zealand psychology. Wellington: New Zealand Psychological Society. Perrewe, P. L., & Zellars, K. L. (1999). An examination of attributions and emotions in the transactional approach to the organizational stress process. Journal of Organizational Behavior, 20, 739-752. Powell, D. (1990). A study of police discretion in six southern cities. Journal of Police Science and Administration, 17, 1-7. Regoli, R., Crank, J., & Culbertson, R. (1989). Police cynicism, job satisfaction and work relations of police chiefs: An assessment of the influence of department size. Sociological Focus, 22(3), 161-171. Schaubroeck, J. (1999). Should the subjective be the objective? On studying mental process, coping behavior, and actual exposures in organizational stress research. Journal of Organizational Behavior, 20, 753760. Sewell, J. D. (1981). Police stress. FBI Law Enforcement Bulletin, 50, 7-11. Silbert, M. H. (1982). Job stress and burnout of new police officers. Police Chief, 49, 46-88. Singleton, G. W., & Teahan, J. (1978). Effects of job-related stress on the physical and psychological adjustment of police officers. Journal of Police Science and Administration, 6, 355-361. Speilberger, C. D., Westberry, L. G., Grier, K. S., & Greenfield, G. (1981). Police stress survey: Sources of stress in law enforcement. Tampa, FL: Human Resources Institute. Stevenson, T. M. (1988). Stress among police officers: Burnout and its correlates, California School of Professional Psychology. Storch, J. E., & Panzarella, R. (1996). Police stress: State-trait anxiety in relation to occupational and personal stressors. Journal of Criminal Justice, 24, 99-107. Stotland, E., & Pendleton, M. (1989). Workload, stress, and strain among police officers. Behavioral Medicine, 15, 5-18. Teahan, J. E. (1975). A longitudinal study of attitude shifts among Black and White police officers. Journal of Social Issues, 31, 47-55. Terry, D. J., & Jimmieson, N. L. (1999). Work control and employee well-being: A decade review. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organizational psychology (pp. 95148). Chichester, UK: Wiley. Terry, W. C. (1985). Police stress as a professional self-image. Journal of Criminal Justice, 13, 501-512. Violanti, J. M., & Aron, F. (1993). Sources of police stressors, job attitudes and psychological distress. Psychological Reports, 72, 899-904. Violanti, J. M., & Aron, F. (1994). Ranking police stressors. Psychological Reports, 75, 824-826. Walker, M. (1997). Conceptual and methodological issues in the investigation of occupational stress: A case study of police officers deployed on body recovery duty. Policing and Society, 7, 1-17. Washington, B. (1981). Stress and the female officer. In L. Territo & H. J. Vetter (Eds.), Stress and police personnel (pp. 142-147). Needham Heights, MA: Allyn and Bacon. Wexler, J. G., & Logan, D. D. (1983). Sources of stress among women police officers. Journal of Police Science and Administration, 11, 46-53. White, S. E., & Marino, K. E. (1983). Job attitudes and police stress: An exploration study of causation. Journal of Police Science and Administration, 11, 264-274. Wright, K. N., & Saylor, W. G. (1991). Male and female employees perceptions of prison work: Is there a difference? Justice Quarterly, 8, 508-524. Merry Morash, Ph.D., is a professor in the School of Criminal Justice at Michigan State University. Robin Haarr, Ph.D., is a Fulbright Senior Specialist currently in Tajikistan working as an international consultant for UNICEF on violence against children, the Swiss Agency for Development and Cooperation on violence against women, and the Organization for Security & Co-operation of Europe on gender mainstreaming in the Government of Tajikistan. Dae-Hoon Kwak is a graduate student in the School of Criminal Justice.